Case study · Outbound rebuild · June to September 2026
An outbound rebuild.
In June 2026 we took over Deligatr’s outbound estate and rebuilt it. Bounce fell from 17% to 2.3%. The best campaign since replied at 11.2%, with 79% of those replies positive and no bounces.
Facts as of 13 September 2026
17.0% → 2.3%
bounce per contact, worst account, before and after
154 of 904 → 101 of 4,465
0
bounces on the best campaign
348 sends to 125 people
11.2%
reply rate on the best campaign
14 of 125 people
79%
of those replies positive
11 of 14, Smartlead’s category
Deligatr (deligatr.ai) · free CRM with lead generation built in, for professional-services firms · Engagement: audit and rebuild, June to September 2026
Before
Deligatr sells a free CRM with lead generation built in, and runs cold email and LinkedIn for professional-services firms under its own white label. Four GTM engineers before us had sold the dream and not delivered. We took over six live accounts in June 2026 after months in which work had been billed but not shipped, and nothing was visible after a client’s onboarding call.
The audit put numbers on it. Zero calendar bookings across all six accounts. Replies sat in sequencer inboxes and never reached the CRM. Every campaign had first sent 18 to 27 days after its mailboxes were created, against a 30-day warm-up standard. Bounce ran at 15.0% and 16.4% on the two worst accounts.
What we did
- 01
Audit first. Every account read before anything changed: 13 issues, a seven-item repair plan, and a baseline to measure against.
- 02
One infrastructure standard. Dedicated domains, inboxes plus a reserve, a full 30-day warm-up, and matched sender names across email and LinkedIn.
- 03
Every reply into the CRM. A positive reply on either channel now creates a contact and an opportunity with the thread attached.
- 04
Lists and mailboxes held to a standard. Company-first lists with a 3% projected-bounce cap. A mailbox audit found 14 blocklisted sending domains and cut the 18 worst inboxes by 88% the same afternoon. It now runs weekly and fails if a healthy mailbox drifts.
Outcome
| Bounce, worst account, per contact | 17.0% (154 of 904) → 2.3% (101 of 4,465) |
|---|---|
| Bounce, second account, per contact | 15.2% (116 of 763) → 2.8% (44 of 1,589) |
| Bounce per send after the relaunch, 28 June | 1.0%, 0.6% and 1.9% on the three relaunched accounts, from 15.0% and 16.4% |
| Best campaign | 125 people · 348 sends · 14 replies (11.2%) · 11 positive (79% of replies) · 0 bounces |
| Second campaign | 653 people · 663 sends · 19 replies (2.9%) · 10 positive (53% of replies) |
| Whole estate | 44,809 emails to 29,828 people · 242 replies · every reply read by a person |
How we count. People contacted means distinct people who received at least one email. Reply rate is replies divided by people contacted. Positive is Smartlead’s own category, as the screenshots show; we also read every reply ourselves. Bounce is bounced sends divided by sends unless marked per contact. Every percentage carries its count.
17.0% → 2.3%
bounce per contact, worst account, before and after
154 of 904 → 101 of 4,465
0
bounces on the best campaign
348 sends to 125 people
11.2%
reply rate on the best campaign
14 of 125 people
79%
of those replies positive
11 of 14, Smartlead’s category
Deligatr (deligatr.ai) · free CRM with lead generation built in, for professional-services firms · Engagement: audit and rebuild, June to September 2026
Before
Deligatr sells a free CRM with lead generation built in, and runs cold email and LinkedIn for professional-services firms under its own white label. Four GTM engineers before us had sold the dream and not delivered. We took over six live accounts in June 2026 after months in which work had been billed but not shipped, and nothing was visible after a client’s onboarding call.
The audit put numbers on it. Zero calendar bookings across all six accounts. Replies sat in sequencer inboxes and never reached the CRM. Every campaign had first sent 18 to 27 days after its mailboxes were created, against a 30-day warm-up standard. Bounce ran at 15.0% and 16.4% on the two worst accounts.
What we did
- 01
Audit first. Every account read before anything changed: 13 issues, a seven-item repair plan, and a baseline to measure against.
- 02
One infrastructure standard. Dedicated domains, inboxes plus a reserve, a full 30-day warm-up, and matched sender names across email and LinkedIn.
- 03
Every reply into the CRM. A positive reply on either channel now creates a contact and an opportunity with the thread attached.
- 04
Lists and mailboxes held to a standard. Company-first lists with a 3% projected-bounce cap. A mailbox audit found 14 blocklisted sending domains and cut the 18 worst inboxes by 88% the same afternoon. It now runs weekly and fails if a healthy mailbox drifts.
Outcome
| Bounce, worst account, per contact | 17.0% (154 of 904) → 2.3% (101 of 4,465) |
|---|---|
| Bounce, second account, per contact | 15.2% (116 of 763) → 2.8% (44 of 1,589) |
| Bounce per send after the relaunch, 28 June | 1.0%, 0.6% and 1.9% on the three relaunched accounts, from 15.0% and 16.4% |
| Best campaign | 125 people · 348 sends · 14 replies (11.2%) · 11 positive (79% of replies) · 0 bounces |
| Second campaign | 653 people · 663 sends · 19 replies (2.9%) · 10 positive (53% of replies) |
| Whole estate | 44,809 emails to 29,828 people · 242 replies · every reply read by a person |
How we count. People contacted means distinct people who received at least one email. Reply rate is replies divided by people contacted. Positive is Smartlead’s own category, as the screenshots show; we also read every reply ourselves. Bounce is bounced sends divided by sends unless marked per contact. Every percentage carries its count.
Proof
Proof
From the client
I had my fingers burned 4 times before I started working with Charlie. In an era where everyone sounds like they know what they are doing using AI, Charlie actually does know what he is doing. If you want to build an outbound infrastructure that works, then do not even think twice. The only reason I could give for not using him is that I would want him all for myself.
Jody Pattison, Principal, Deligatr. Public review, June 2026.
Results vary by market, list and offer. These figures come from one engagement, counted as described above, and are not a forecast for yours.
Written by
Charlie Brown, founder of Groundswell.
Who this is for
- Agencies running outbound for several clients
- Founders who inherited someone else’s sending estate
- Anyone whose sequencer shows replies the CRM does not
Stack on this build
- Smartlead
- HeyReach
- GoHighLevel
- Clay
- n8n
- GitHub Actions
Results vary by market, list and offer. These figures come from one engagement, counted as described above, and are not a forecast for yours.
Written by
Charlie Brown, founder of Groundswell.
Who this is for
- Agencies running outbound for several clients
- Founders who inherited someone else’s sending estate
- Anyone whose sequencer shows replies the CRM does not
Stack on this build
- Smartlead
- HeyReach
- GoHighLevel
- Clay
- n8n
- GitHub Actions
Talk it through
Running outbound for clients and not sure what it produces? Thirty minutes, no pitch.
Talk it through
Running outbound for clients and not sure what it produces? Thirty minutes, no pitch.
Talk it through
Running outbound for clients and not sure what it produces? Thirty minutes, no pitch.